Library screening using ELISA Screening of a PS-SCL for inhibition of mab binding to antigen by ELISA 65

Transcription

1 Protocol list Manual and automated synthesis Manual synthesis of a resin-bound pentapeptide library 6 Automated construction of an organic library using split synthesis: 3-amino-5-hydroxybenzoic acid as a core structure 10 Synthetic strategies for library construction Portioning-mixing synthesis of a library encoded with molecular tags 18 Synthesis of a radiofrequency encoded library 24 Synthesis of the peptide library Preparation of the amino acid solutions for a heptapeptide library synthesis 35 Synthesis of a linear peptide library with 19 eukaryotic amino acids (cysteine excluded) 36 Synthesis of the disulfide cyclic peptide library on solid phase 38 Synthesis of Dpr-Dpr cyclic peptide (oxime bond) library 40 Synthesis of cyclic peptide libraries using on-resin cyclization between Lys and Glu side-chains 41 Library screening Enzyme-linked colorimetric assay 43 Unlabelled ligate detected with an enzyme-linked secondary antibody system 45 Cross-screening the library with enzyme-linked colorimetric and radiolabelled assays 46 Determination of peptide substrate motifs for protein kinases 47 Mixture-based synthetic combinatorial libraries Synthesis of a tripeptide PS-SCL 55 Cleavage and extraction of a PS-SCL 55 Alkylation of a resin-bound peptide PS-SCL 57 Reduction of a resin-bound PS-SCL 59 Formation of a bicyclic guanidine PS-SCL 61 Library screening using ELISA Screening of a PS-SCL for inhibition of mab binding to antigen by ELISA 65 Identification of T cell-specific ligands Screening of a PS-SCL to identify CD4 + or CD8 + T cell ligands 68 Identification of antimicrobial and antifungal compounds Screening of a PS-SCL for identification of antibacterial compounds 69 xxiii

2 PROTOCOL LIST Library screening using a radioreceptor assay Screening of a PS-SCL in a radioreceptor assay 70 Identification of enzyme inhibitors using PS-SCLs Screening of a PS-SCL in an -glucosidase inhibition assay 72 Applications in combinatorial synthesis Application of R f encoded MicroKans in combinatorial synthesis 82 Performing reactions with MicroKans: Wittig alkene formation 86 Performing reactions with MicroKans: cross olefin metathesis 87 Performing reactions on MicroTubes: reductive amination 90 Oligonucleotide synthesis on bar coded microreactors 92 SPOT synthesis of peptides on continuous cellulose surfaces Preparation of amino-functionalized cellulose membranes (ester linkage) 96 Preparation of ester-free amino-functionalized cellulose membranes 97 Determining the loading of an amino-functionalized cellulose membrane 98 Definition of the SPOTS 99 The SPOT synthesis coupling cycle 100 Cleavage of side-chain protecting groups 100 Cleavage of peptides from the cellulose membrane 101 Synthesis of PNA arrays using the SPOT technique PNA assembly on cellulose membranes 104 Preparation of stable polymeric membranes for SPOT synthesis of organic compound libraries Preparation of PP-g-P(PEGMA) membranes 107 Attachment of linkers to PP-g-P(PEGMA) membranes 108 Preparation of PP-g-PAA membranes 109 Synthesis of PP-g-P(AmPEGMAm) membranes 110 Synthesis of a tripeptoid using the SPOT synthesis method 110 Synthesis of a triazine derivative via SPOT synthesis method 111 Linkers and anchors in SPOS Attachment of carboxylic acids to PS/DVB-HMBA resin 133 Attachment of nucleophiles to 2-chlortrityl chloride PS/1% DVB resin 134 Attachment of carboxylic acids to PS/DVB-Wang resin 134 Quantification of coupling sites on solid support by Fmoc (fluorenylmethoxycarbonyl) cleavage 135 Analytical methods in reaction optimization Single bead FTIR measurement 264 Beam condenser FTIR measurement 266 General procedure for gel phase NMR 267 General procedure for MAS NMR spectroscopy 268 Quantitative determination of aldehyde and ketone groups on resins 271 Quality control of libraries from parallel synthesis High-throughput flow injection analysis (FIA) MS 273 The screening of pooled combinatorial libraries Enzyme-linked immunosorbent assay for identifying ligands 279 Decoding the ligands by single bead analysis using matrix-assisted laser desorption and ionization mass spectrometry 280 xxiv

3 PROTOCOL LIST Isocyanide chemistry Preparation of alkyl-isocyanides from N-alkylformamides using diphosgene 289 Preparation of isocyanides from the corresponding formamides using POCl 3 and diisopropylamine 289 MCR chemistry N,N-phthalyl-glycil-N -benzyl-valine-tert-butylamide Stereoselective U-4CRs and their applications in the synthesis of α-amino acids, peptides, and related compounds Preparation of -amino acid and peptide derivatives by the stereoselective U-4CR using amino sugar 48 as a chiral auxiliary 294 Procedure for the synthesis of peptide derivative General procedure for the preparation of -lactam Multi-component reactions of five and more reactants Synthesis of bicyclic 1,3-diketopiperazine derivative 61 from valine methylester 58, levulinic acid 57, and methyl isocyanide The preparation of a thiazole derivative by a one-pot reaction of an U-4CR and a secondary reaction 299 Cyclic anhydride chemistry with extractive purification Synthesis of N-Boc-iminodiacetic acid 308 First diversification of iminodiacetic acid 309 Second diversification of iminodiacetic acid 310 Third diversification of iminodiacetic acid 312 Higher order libraries Dimerization of iminodiacetic acid diamides via dicarboxylic acid coupling 315 Olefin metathesis dimerization coupling 316 Tetramerization of iminodiacetic acid diamides by sequential diacid couplings 318 Olefin metathesis tetramerizations 320 General aspects of fluorous chemistry Preparation of FRP silica gel 332 Fluorous tin chemistry Synthesis of tris(2-perfluorohexylethyl)phenyltin (1), tris(2-perfluorohexylethyl)tin bromide (2), and tris(2-perfluorohexylethyl)tin hydride (3) 334 Parallel experiment with tin hydride 3 under catalytic conditions (9-component library) 336 Synthesis and reaction of tin azide Rapid fluorous Stille coupling reactions with microwave heating 339 Synthesis of propylene-spaced fluorous allyltin reagents 341 Parallel allylation with separation by FRP silica solid phase extraction 343 Fluorous synthesis Fluorous isoxazoline synthesis without intermediate purification 345 Fluorous Ugi reaction 346 Fluorous quenching (scavenging) Tin hydride quench of an excess alkene by hydrostannation 348 Fluorous amine quench for 3 3 library 350 Supports for COSMOS Synthesis of soluble support PEG4-Rink xxv

4 PROTOCOL LIST Organic synthesis in COSMOS Synthesis of a trisubstituted guanidine using a soluble support 362 Size-based purification of homogeneous reaction products in COSMOS Packing a SEC column with Bio-Beads S-X1 366 Assembling a system for sequential product isolation by SEC 368 Product purification by SEC 369 Applications of automated synthesis systems Use of an automated system for the synthesis of 3-(imidazo[1,2-b]pyridazin- 6-yl)thiopropanesulfonamide 382 Use of an automated system for the synthesis of a dipeptide derivative, Boc-Glu(OcHex)-Leu-OBn 385 Use of an automated system for investigating the selectivity of O-acylation of 3-(1-hydroxyethyl)-4-acetoxyazetidin-2-one 389 Applications of automated synthesis workstations Use of an automated workstation for the multigram scale synthesis of an intermediate peptide derivative, Boc-D-Ala-OBn 392 Use of an automated workstation for the multigram scale synthesis of a tetrapeptide derivative, Boc-Lys(Z)-D-Ala-Tyr(Bn)-D-Ala-OBn 395 Preparation of 9-cyanophenanthrene using slow addition of a powdered supported reagent 397 Batch preparation of Boc-Glu(OcHex)-D-Ala-OH using catalytic hydrogen transfer 399 Optimization of the catalyst processing conditions Synthesizing and testing individual anode catalysts 405 Preparation of electrode arrays Electroplating gold onto the stainless steel screen 409 Printing an array of electrocatalysts 411 Optimizing the screening and testing conditions Screening anode electrocatalyst arrays 413 Tetrahydroquinoline library Preparation of polyallylscandium trifylamide ditriflate 424 Tetrahydroquinoline synthesis 425 β-amino ketone and ester library -Amino ketone and ester synthesis 428 α-amino nitrile library -Amino nitrile synthesis 429 Catalyst discovery (optimization of reactivity) Solid phase synthesis of peptidyl Schiff base ligands 437 Indexed grid 439 Solid phase ligands in the epoxide ring opening reaction 440 Catalyst optimization (optimization of selectivity) Cleavage of peptidyl Schiff base ligands prepared on resin 445 Testing of the solution phase ligands in the epoxide ring opening reaction 445 Positional scanning of Strecker reaction on solid support 447 Preparative Ti catalysed addition of TMSCN to imines 450 Conversion of amino nitriles to BOC-protected amino acids 450 xxvi

5 Chapter 2 One-bead one-compound combinatorial library method Gang Liu and Kit S. Lam Department of Internal Medicine, University of California, UC Davis Cancer Center, 4501 X Street, Sacramento, CA 95817, USA 1 Introduction The one-bead one-compound combinatorial library method introduced by Lam et al. (1) involves: (a) The generation of a large library of spatially separable synthetic compounds. (b) The screening of this library of compounds for a specific biological, chemical, or physical property. (c) The isolation of the positive compounds for structural determination. The one-bead one-compound combinatorial library was prepared by a split synthesis method (1 3, see also Chapter 1) such that each solid phase bead displays only one chemical entity although there are approximately copies of the same compound on a 100 m bead. This bead-supported library is then assayed for a specific biological activity using on-bead binding (1), or functional assays (4). Alternatively, the compound on each bead is released from the solid support via a cleavable linker for subsequent solution phase assays (5). In either solid phase or solution phase assays, the positive beads are then identified, physically isolated, and subjected to structural determination. If the compound on the bead is a peptide, each individual positive bead is inserted into a cartridge of an automatic protein sequencer for sequence analysis. Alternatively, the peptide or non-peptide compound on each bead is released chemically or photochemically and the releasate is analysed by mass spectrometry. If the structure of the compound cannot be easily determined by mass spectrometry, one can use chemical encoding (6 8) or radiofrequency tagging (9, Chapter 4) strategies. Synthetic combinatorial chemistry was first applied to peptides (1, 2, 10). However, numerous papers on the development of methods to synthesize small 33

6 GANG LIU AND KIT S. LAM molecule combinatorial libraries were reported in the literature over the past six years (11 14). A detailed account of the various synthetic methods, screening strategies, or structure determination methods are beyond the scope of this chapter. Please refer to ref. 14 for a recent comprehensive review of the onebead one-compound combinatorial library method. In this chapter we describe selected protocols for the preparation of linear and cyclic peptide libraries, and various on-bead binding as well as functional screening assays that have been employed in our laboratory. 2 Synthesis of the peptide library 2.1 Preparation of the amino acid solutions Stock solutions of N -Fmoc-AAs (N -fluorenylmethoxycarbonyl amino acids) with HOBt (N-hydroxybenzotriazole) in DMF (N,N-dimethylformamide) or NMP (N-methylpyrrolidine) can be stored at 4 C for up to ten days. Three fold excess of N -Fmoc-AAs are used in each coupling cycle to ensure completion of the peptide coupling reaction. The following equations are used for the calculation of the necessary reagents. (MW aa )(3L)(g)(N) W aa (1) X W HOBt (135.1)(3L)(g)(N) (2) V HOBt (0.71)(V )(X)(N) (3) (126.1)(3L)(g) V DIC (4) X(0.806) MW aa molecular weight of the protected amino acid; L loading of the resin in mmol/g; g mass of the resin; N total number of amino acid residues in the peptide sequence; X total number of amino acids used in each coupling step; V volume of amino acid/hobt solution to be added to each reaction vessel (for a reaction vial containing 100 mg TentaGel resin, we generally use 0.5 ml amino acid/hobt solution); specific gravity of DIC (N,N -diisopropylcarbodiimide); molecular weight of HOBt; molecular weight of DIC. Equation 1 is used to calculate the mass of each amino acid required for the library synthesis. Equation 2 is used to calculate the amount of HOBt necessary for the whole library. Equation 3 is used to calculate the volume of HOBt solution needed. Equation 4 is used to determine the volume of DIC required for each amino acid coupling reaction. 34

7 ONE-BEAD ONE-COMPOUND COMBINATORIAL LIBRARY METHOD Protocol 1 Preparation of the amino acid solutions for a heptapeptide library synthesis Reagents The following N -Fmoc-AAs with the appropriate side-chain protecting groups are used in our library synthesis: N -Fmoc- Arg(Pmc)-OH, N -Fmoc-Asn(Trt)-OH, N - Fmoc-Gln(Trt)-OH, N -Fmoc-Asp(Ot-Bu)- OH, N -Fmoc-Glu(Ot-Bu)-OH, N -Fmoc- His(Trt)-OH, N -Fmoc-Lys(Boc)-OH, N - Fmoc-Ser(t-Bu)-OH, N -Fmoc-Thr(t-Bu)-OH, N -Fmoc-Trp(Boc)-OH, N -Fmoc-Tyr(t-Bu)- OH The other amino acids without side-chain protecting group are: N -Fmoc-Ala, N -Fmoc-Phe, N -Fmoc-Gly, N -Fmoc-Ile, N -Fmoc-Leu, N -Fmoc-Met, N -Fmoc-Pro, N -Fmoc-Val HOBt-H 2 O HPLC (high performance liquid chromatography) grade DMF or NMP 2.0 g TentaGel S NH 2 resin (0.27 mmol/g) Method 1 Use Equations 1 4 to calculate the amount of reagents necessary for the synthesis of a heptapeptide library (2.0 g resin) using 19 eukaryotic a amino acids. W aa MW aa / MW aa mg. W HOBt mg. V HOBt ml. b V DIC / l. 2 Pre-weigh the solid reagents into a suitable container (e.g. graduated polypropylene vials). 3 Dissolve 1532 mg HOBt in DMF until the final volume is 46.6 ml (V HOBt ). 4 Distribute the HOBt solution (2.45 ml each) into each of the 19 amino acid containers and add HPLC grade DMF until the final volume is (V ) (N) or 3.5 ml. Try to dissolve the amino acid by vortexing. If it does not dissolve totally, try sonication. c a Naturally occurring amino acids. b 0.5 ml of amino acid and HOBt solution (V ) will be added into each reaction vessel at each coupling step. c If suspension does not dissolve totally after sonication ensure that it is well mixed prior to pipetting into the reaction vessel. 2.2 Synthesis of a linear peptide library on solid phase using TentaGel resin with 19 eukaryotic amino acids (cysteine excluded) Lam et al. (14) have reviewed the choice of solid supports for the one-bead onecompound combinatorial chemistry methodology. TentaGel resin was generally used as the solid phase support. It consists of polystyrene matrix with a grafted polyoxyethylene (POE) linker and a functional group (amino or hydroxyl) at the 35

8 GANG LIU AND KIT S. LAM end of this linker. This resin is fully compatible with both organic and aqueous conditions. The POE linker is particularly suitable for many biological assay conditions. In principle, either the t-boc/bn (tertio-butyloxycarbonyl/benzyl) or Fmoc/t-But (fluorenylmethoxycarbonyl/tertio-butyl) amino acid protection strategies could be used for library synthesis on TentaGel. For the t-boc/bn chemistry, HF (hydrofluoridric acid) is needed to cleave off the side-chain protecting groups (CAUTION: HF is highly toxic and requires a special apparatus). Since it partially degrades the POE chain on the resin, we recommend the Fmoc/t-But strategy for the synthesis of peptide libraries on TentaGel. Many of the commercially available activating reagents can be used for solid phase peptide library synthesis. These include DIC, BOP (benzotriazol-1-yl-oxy-tris (dimethylamine)-phosphonium hexafluorophosphate), HBTU (O-benzotriazole- N,N,N,N -tetramethyluronium hexafluorophosphate), TBTU (O-benzotriazol-1-yl- N,N,N,N -tetramethyluronium tetrafluoroborate), PyBOP (benzotriozol-1-yl-oxytris-pyrrolidinophosphonium hexafluorophosphate), and PyBroP (bromo-trispyrrolidine-phosphonium hexafluorophosphate). However, DCC (N,N -dicyclohexylcarbodiimide) should be avoided because it will produce an insoluble DCU (dicyclohexylurea) side-product. An additive reagent to generate an active ester is needed to reduce racemization of the N -protected amino acids. HOBt and Pfp (pentafluorophenyl) esters are the most efficient. We usually use HOBt for this purpose. When BOP, PyBOP, TBTU, HBTU, or PyBroP are used as the activating reagents and HOBt as an additive, a base must be added in situ to neutralize the acidic side-product that may deprotect the acid-labile amino acid side-chain protecting groups such as Trt (trityl) and t-bu. We recommend DIEA (N,N-diisopopylethylamine), NMM (N-methylmorpholine), or TEA (triethylamine). In the following protocol we describe a procedure using DIC as an activating reagent and HOBt as an additive to synthesize a peptide library. Protocol 2 Synthesis of a linear peptide library with 19 eukaryotic amino acids (cysteine excluded) Reagents Protected amino acid stock solutions (Protocol 1) 25% piperidine/dmf (v/v) TFA (trifluoroacetic acid) Kaiser test reagents: solution A, 5% ninhydrin in ethanol (w/v); solution B, 80% crystalline phenol in ethanol (w/v); solution C, 2% M aqueous solution of potassium cyanide in pyridine (v/v) DIC TentaGel S NH 2 resin ( m), loading: 0.27 mmol/g 1,2-ethanedithiol (EDT) Cleavage solution: phenol/thioanisole/h 2 O/EDT/TFA (0.75:0.5:0.5:0.25:10, w/v/v/v/v) PBS (phosphate-buffered saline): 137 mm NaCl, 2.68 mm KCl, 8.0 mm Na 2 HPO 4, 1.47 mm KH 2 PO 4 ph

9 ONE-BEAD ONE-COMPOUND COMBINATORIAL LIBRARY METHOD Protocol 2 continued Method 1 Weigh out 2.0 g TentaGel S NH 2 resin (0.27 mmol/g) and swell it in HPLC grade DMF overnight. 2 Distribute the beads equally into 19 reaction vessels. a Allow the resin to settle for 20 min and remove the DMF carefully with a disposable polyethylene pipette down to the surface of the resin. 3 Add 0.5 ml amino acid/hobt solution (Protocol 1, step 4). b 4 Add DIC to each reaction vessel as calculated by Equation 4. 5 Couple the amino acid for 1 h at room temperature with gentle shaking. 6 Transfer a small sample of the resin solution from each reaction vessel into tiny test-tubes. 7 Wash each mixture in the tubes with ethanol once and then perform the Kaiser test. c,d 8 After all the couplings are completed, combine the resins and transfer them to a siliconized column with a frit at the bottom. 9 Wash with DMF (5 10 ml). e 10 Remove the N -Fmoc protecting group by incubation twice in 10 ml of 25% piperidine/dmf for 15 min each. 11 Wash the resin again with DMF (6 5 ml), methanol (2 5 ml), and DMF (6 5 ml). f 12 Repeat the cycle of resin distribution, amino acid coupling, Kaiser test, resin mixing, resin washing, Fmoc deprotection, and resin washing again according to steps 2 11 until amino acid assembly is completed. 13 Remove the N -Fmoc protecting group as described in steps 10 and 11 prior to sidechain deprotection. 14 Remove the side-chain protecting groups by adding 30 ml cleavage solution to g bead-supported library for 2.5 h at room temperature with gentle mixing. 15 Drain the deprotection solution, and wash the bead-supported library with DMF (5 5 ml), methanol (5 5 ml), DCM (5 5 ml), and DMF (5 5 ml). 16 Wash the bead-supported library with 30% H 2 O/DMF, 60% H 2 O/DMF, 100% H 2 O (1 10 ml each), and then 50 mm PBS buffer ph 7.4 (10 10 ml). 17 Store the peptide-bound resin in 0.05% sodium azide/pbs at 4 C. a Polypropylene or polyethylene vials with tight-sealing cap. b Use HPLC grade DMF for the coupling reaction. c Transfer a minimum amount of resin beads to a small glass test-tube, wash the beads with ethanol. Add one drop each of solutions A, B, and C to the tube containing the beads. Heat the mixture to 100 C and maintain the heat for 5 min. Blue coloured beads and solution (positive) 37

10 GANG LIU AND KIT S. LAM Protocol 2 continued indicates the presence of free amino groups on the resin, i.e. incomplete coupling. Colourless resin beads indicate complete coupling. For those vessels with incomplete coupling reaction, either prolong their reaction time or remove the supernatant and perform a repeat coupling by adding fresh amino acid/hobt solution and DIC. d For proline, a secondary amine, the Kaiser test will turn brown instead of blue. e We suggest using technical grade DMF for washing since it is inexpensive. f Use HPLC grade DMF in the last three washings. 2.3 Synthesis of disulfide cyclic peptide library Careful choice of oxidation methods and protecting groups for cysteine are important for subsequent disulfide formation. There are three main methods to cyclize a peptide using disulfide bond formation in solution or on solid support: (a) Air oxidation. (b) DMSO oxidation of the free thiol ( SH) groups to form a disulfide bond ( S S ) in polar solvents such as methanol/water or acetic acid (AcOH)/water. (c) Iodine (I 2 ) oxidation of the Cys(S-Trt)/Cys(Acm) directly to form disulfide bond in methanol or AcOH. We recommend using method (b) for free thiol oxidation and method (c) for the protected Cys(S-Trt)/Cys(S-Acm) oxidation. In the following protocol, we describe a procedure for the synthesis of disulfide peptide library. During library synthesis, Cys(S-Trt)-OH is added at the first and last residual of the library. The Trt protecting group is removed with TFA during the side-chain deprotection step. Cyclization is accomplished by disulfide formation using DMSO (method b). Since 1,2-ethanedithiol (EDT) can reduce disulfide bonds during the oxidation step, it must be replaced by triisopropylsilane (TIS) as the scavenger in the cleavage protocol. Protocol 3 Synthesis of the disulfide cyclic peptide library on solid phase Reagents N -Fmoc-Cys(S-Trt)-OH TIS Cleavage solution: phenol/thioanisole/h 2 O/TIS/TFA (0.75:0.5:0.5:0.25:10, w/v/v/v/v) Ellman s reagent: 4.0 mg 5,5 -dithio-bis(2- nitrobenzoic acid) in 1.0 ml of 20 mm sodium phosphate buffer ph 8.0 Concentrated aqueous ammonia Reagents for peptide synthesis (Protocol 2) 38

11 ONE-BEAD ONE-COMPOUND COMBINATORIAL LIBRARY METHOD Protocol 3 continued Method 1 Weigh out 2.0 g TentaGel S NH 2 resin (0.27 mmol/g) and swell it in HPLC grade DMF overnight. 2 Dissolve 3.0 Equiv. of N -Fmoc-Cys(S-Trt)-OH (949 mg) based on the loading of resin in 15 ml DMF. 3 Add 3.0 Equiv. of HOBt (219 mg) and 3.0 Equiv. of DIC (255 l) to this solution. Then add the final mixture to the resin and shake overnight. 4 After the coupling is completed (Kaiser test must be negative), remove the N -Fmoc protecting group by incubation twice in 10 ml of 25% piperidine/dmf for 15 min each. 5 Wash the resin as in Protocol 2, step To assemble the peptide library, follow the steps outlined in Protocol 2, steps Couple cysteine at the desired position of the peptide library by adding 3.0 Equiv. of N -Fmoc-Cys(S-Trt), HOBt, and DIC as in steps 2 and 3. 8 After the peptide library assembly is completed, remove the N -Fmoc protecting group according to step 4. 9 Remove the side-chain protecting groups by adding 30 ml of the cleavage solution to g of bead-supported library for 2.5 h at room temperature with gentle shaking. 10 Drain the cleavage solution, and wash the bead-supported library with 5 5 ml each of DMF, methanol, DCM, and DMF again. 11 Wash the bead-supported library with 10 ml each of 30% H 2 O/DMF, 60% H 2 O/DMF, H 2 O, and finally test for unreacted free thiol groups on the resin using Ellman s reagent. a 12 Oxidize the bead-supported library in 1 litre of H 2 O/HOAc/DMSO b (75:5:20) with gentle stirring in a 2 litre flask for 48 h, until the Ellman test is negative. 13 Wash the beads thoroughly with H 2 O, followed by PBS. 14 Store the resin-bound library in 0.05% sodium azide/pbs at 4 C. a In this step, Ellman test must be positive, the yellow colour indicates that 5,5 -dithio-bis(2- nitrobenzoic acid) has reacted with free thiol group on the resin to form yellow anion product. b The ph is adjusted with saturated aqueous ammonium hydroxide to 6.0 before DMSO is added. 2.4 Synthesis of Dpr-Dpr cyclic peptide (oxime bond) library Wahl and Mutter (15) first introduced two new unnatural amino acids (N -Fmoc- Dpr(t-Boc-Aoa)-OH and N -Fmoc-Dpr(Boc-Ser(Ot-Bu)-OH) as building blocks for the solution phase cyclization of peptides via oxime bond formation. The oxime bond is stable to proteolysis and represents an excellent alternative approach for constraining peptide conformation. Here we describe a method for on-resin cyclization of a peptide library using the same approach. 39

12 GANG LIU AND KIT S. LAM Protocol 4 Synthesis of Dpr-Dpr cyclic peptide (oxime bond) library Reagents N -Fmoc-Dpr(t-Boc-Aoa)-OH and N -Fmoc- Dpr(Boc-Ser(Ot-Bu))-OH 115 mg NaIO 4 in 25 ml of 50% H 2 O/DMF Reagents for peptide synthesis (Protocol 2) Method 1 Weigh out 2.0 g TentaGel S NH 2 resin (0.27 mmol/g) and swell it in HPLC grade DMF overnight. 2 Dissolve 2.0 Equiv. of N -Fmoc-Dpr(Boc-Ser(Ot-Bu))-OH (615 mg, based on the loading of the resin) in 15 ml DMF. Add 2.0 Equiv. of HOBt (146 mg) and 2.0 Equiv. of DIC (170 l) into this solution. Finally, add the beads and shake overnight. 3 After the Kaiser test is negative, remove the N -Fmoc protecting group then wash the resin as in Protocol 2, steps 10 and Assemble the peptide library according to Protocol 2, steps Couple 2.0 Equiv. of N -Fmoc-Dpr(Boc-Aoa) (539 mg) with the addition of reagents HOBt (146 mg) and DIC (170 l). 6 Remove the side-chain protecting groups with 30 ml of cleavage solution as in Protocol 2. 7 Wash the resin thoroughly with DMF and then add twice 2.0 Equiv. of NaIO 4 (115 mg) in 25 ml of 50% H 2 O/DMF for 10 min each to convert the serinyl to a glyoxylyl group. The beads turn red. 8 Wash the bead-supported library with 50% H 2 O/DMF (5 10 ml), and DMF (3 10 ml). 9 Remove the last N -Fmoc protecting group on the N-terminus of the peptide library by incubating twice in 10 ml of 25% piperidine/dmf for 15 min each. The colour of the beads returns to light yellow. 10 Wash the beads with DMF (5 10 ml), 30% H 2 O/DMF (1 10 ml), 60% H 2 O/DMF (1 10 ml), 100% H 2 O (1 10 ml), and 1 PBS (10 10 ml). 11 Store the peptide-bound resin in 0.05% sodium azide/pbs at 4 C. 2.5 On-resin synthesis of a cyclic peptide library using Lys and Glu side-chains On-resin amide bond formation between the amino and carboxyl group sidechains of Lys and Glu, respectively, is a useful approach to cyclize peptides while still attached to the solid support. Several protecting group strategies were developed for this purpose, which include: (a) Head-to-tail cyclization using t-boc-asp-ofm or t-boc-glu-ofm (16), and N - Fmoc-Asp-OAll or N -Fmoc-Glu-OAll (17). 40

13 ONE-BEAD ONE-COMPOUND COMBINATORIAL LIBRARY METHOD (b) Side-chain cyclization using N -Fmoc-Asp(ODmab)-OH or N -Fmoc- Glu(ODmab)-OH with N -Fmoc-Lys(Dde)-OH. (c) Another side-chain based cyclization using t-boc-asp(ofm)-oh or t-boc- Glu(OFm)-OH with t-boc-lys(fmoc)-oh. The Fm group is removed in 20% piperidine/dmf, while the allyl group is cleaved by Pd(PPh 3 ) 4 in CHCl 3 :AcOH:N-methylmorpholine (37:2:1) as described in NovaBiochem Catalog Handbook (18). Both the Dmab and Dde groups are labile to 2% hydrazine in DMF. In the following protocol, we choose N -Fmoc- Lys(Dde)-OH and N -Fmoc-Glu(ODmab)-OH as building blocks for cyclization. These amino acids couple through the -NH 2 group of Lys and -carboxyl group of Glu. Protocol 5 Synthesis of cyclic peptide libraries using on-resin cyclization between Lys and Glu side-chains Reagents N -Fmoc-Lys(Dde)-OH and N -Fmoc- Glu(ODmab)-OH PyBOP 1-Hydroxy-7-azabenzotriazole (HOAt) DIEA 2% NH 2 NH 2 /DMF (v/v) Reagents for peptide synthesis (Protocol 2) Method 1 Weigh out 2.0 g TentaGel S NH 2 resin (0.27 mmol/g) and swell it in HPLC grade DMF overnight. 2 Decrease the loading of the resin using the following method. (a) Mix 0.30 mmol of N -Fmoc-Gly (89 mg), 0.31 mmol of HOBt (42 mg), and 0.31 mmol of DIC (48.5 l) in 15 ml of DMF. (b) Add the mixture to the TentaGel S NH 2 resin and shake gently overnight at room temperature. (c) After washing the resin with DMF and DCM, cap the exposed amino group with 15% acetic anhydride in DCM for 30 min. In this case, the final loading of the down-substituted resin will be 0.15 mmol/g. a 3 Cleave the N -Fmoc protecting group and wash the resin as in Protocol 2, steps 10 and Dissolve 2.0 Equiv. of N -Fmoc-Lys(Dde)-OH (320 mg based on the new loading of the resin) in 15 ml DMF. Add 2.0 Equiv. of HOBt (81 mg) and 2.0 Equiv. of DIC (94 l) to this solution. Finally, add the beads and shake gently overnight. 5 Follow the procedure for assembly of the peptide as described in Protocol 2, steps 2 12, until the desired cyclization position (Glu) is reached. 41

14 GANG LIU AND KIT S. LAM Protocol 5 continued 6 Add 2.0 Equiv. of N -Fmoc-Glu(ODmab) (735 mg), DIC (94 l), and HOBt (81 mg) to the beads and shake gently overnight. The Kaiser test must be negative after this coupling. 7 After washing with DMF (5 10 ml), incubate twice in 80 ml of 2% NH 2 NH 2 /DMF for 3 min each. 8 Wash the beads with DMF (6 10 ml). 9 Perform the cyclization step by adding 5.0 Equiv. of PyBOP and HOAt and 10 Equiv. of DIEA in HPLC grade DMF. Shake gently at room temperature overnight or until the Kaiser test is negative. 10 Remove the N -Fmoc protecting group (Protocol 2, steps 10 and 11). 11 Remove the side-chain protecting groups by incubation in 30 ml of the cleavage solution (Protocol 2, step 14). 12 Wash the beads with DMF (5 10 ml), DCM (5 10 ml), DMF (5 10 ml), 30% H 2 O/DMF (1 10 ml), 60% H 2 O/DMF (1 10 ml), 100% H 2 O (1 10 ml), and finally with PBS buffer (10 10 ml). 13 Store the resin-bound library in 0.05% sodium azide/pbs at 4 C. a An alternate way to decrease the loading of the resin is to first couple 1.0 Equiv. of a mixture of Fmoc-Gly and Ac-Gly in a 1:1 molar ratio to the resin overnight at room temperature, then perform a coupling with 3.0 Equiv of N -Fmoc-Gly for 1 h. The final loading will be half of the original loading of the resin. 3 Library screening The one-bead one-compound combinatorial library method has been adapted for both on-bead solid phase as well as releasable solution phase assay (1, 4, 14). For the on-bead binding assay, the receptor molecular target (ligate) needs to be tagged with a reporter molecule such as an enzyme, a fluorescent probe, or a radionucleide. Alternatively, the ligate is biotinylated and in turn the reporter group will be linked to streptavidin which is used as a secondary reagent to detect the positive beads (19). Another approach is to use tagged antibodies that specifically recognize the ligate to label the positive beads (20). Although purified ligate is generally used for the above screening method, it is not crucial. For instance if a highly specific tagged secondary antibody is used in the assay, the ligate need not be highly purified. However, regardless of which screening method one has decided to use, it is crucial to design strategies such that the positive beads from the first screen are recycled and screened again by an additional (orthogonal) method to ensure that the initial positive beads are not erroneous. Besides using soluble ligates as probes for screening, one may also use intact cells or micro-organisms. We have reported the use of intact tumour cell lines to identify integrin-specific binding peptides (21). In this method, the beadsupported library is mixed with intact cells. Positive beads rosetted by a monolayer of cells are isolated, recycled, and retested for binding to the selected beads in the presence of anti-integrin antibodies. We have also used a similar 42

15 ONE-BEAD ONE-COMPOUND COMBINATORIAL LIBRARY METHOD approach to identify ligands that bind specifically to yeast (Candida albican) cell surface (unpublished results). In addition to on-bead binding assay, we and others have developed screening methods to identify ligands with specific function. These functional assays include identification of peptide substrate motif for protein kinases (4, 22, 23) and proteases (24, 25). In the protein kinase assay, beads with peptide substrate motif are covalently modified by 32 P when the bead-supported library is incubated with a specific protein kinase and [ - 32 P]ATP. The 32 P-labelled beads are then identified by autoradiography and isolated for microsequencing. As for the protease substrate identification, the library is constructed such that there is a quencher at the amino-terminus and a fluorophore at the carboxyl-terminus of the peptide library. The peptide sequence susceptible to proteolysis will be cleaved from the solid phase support thereby releasing the quenching moiety and resulting in an enhancement of the fluorescence of the resin-bound fluorophore. The highly fluorescent beads are then identified and isolated for microsequencing. The one-bead one-compound combinatorial library method has also been adapted to solution phase assay in which the ligand is attached to the bead via a cleavable linker (chemo-sensitive or photo-sensitive). Two general approaches have been used. The in situ approach involves the immobilization of the beads prior to release of the compounds, thus the beads from which the active compounds were released can be spatially traced back and identified (26). The second approach uses microwell plate to partition the beads (one to 500 beads/well) prior to the release of the compounds from the beads. When large numbers of beads per well are used, redistribution of beads from an active well for a second release of compounds using a multi-cleavable linker (5, 28) is required in order to identify a unique bead with the desired activity. In the following section we illustrate the above concepts with a few selected screening methods developed in our laboratories (see also ref. 14). Protocol 6 Enzyme-linked colorimetric assay Equipment and reagents Dissecting microscope Automatic protein sequencer (e.g. Applied Biosystems Model ABI 477A) A bead-supported peptide library with a specific composition PBS: 8.0 mm Na 2 HPO 4, 1.5 mm KH 2 PO 4, 137 mm NaCl, 2.7 mm KCl ph 7.4 TBS: 2.5 mm Tris base, 13.7 mm NaCl, 0.27 mm KCl ph 8.0 BCIP substrate buffer: 1.65 mg 5-bromo-4- chloro-3-indoyl phosphate (BCIP) in 10 ml of 0.1 M Tris base, 0.1 M NaCl, 2.34 mm MgCl 2 ph % aqueous gelatin (w/v) (dissolve with heating in microwave oven) 6.0 M guanidine hydrochloride ph 1.0 Ligate alkaline phosphatase conjugate 43

16 GANG LIU AND KIT S. LAM Protocol 6 continued Method 1 Transfer 1 10 ml of the bead-supported library into a disposable polypropylene column with a polyethylene frit. a Wash the bead-supported library with water (5 10 ml). Block the bead-supported library with 0.05% gelatin (w/v) in water for at least 1 h at room temperature. 2 Wash the bead-supported library with 0.1% Tween/PBS thoroughly (10 10 ml). 3 Incubate the bead-supported library with the ligate alkaline phosphatase conjugate at a suitable concentration in 0.05% gelatin/0.1% Tween/PBS from 1 24 h at room temperature. 4 Wash the bead-supported library thoroughly with 0.1% Tween/PBS (10 10 ml) and then with TBS (10 ml). 5 Add BCIP substrate buffer. 6 Transfer and distribute the bead suspension into polystyrene Petri dishes (10 cm diameter). 7 Develop the colour for h at room temperature. 8 Stop the enzymatic reaction with a few drops of 1.0 M HCl. b 9 With the aid of a light box and a hand-held micropipette, transfer the turquoise beads into a small Petri dish. c 10 Under the dissecting microscope, pick up the deep coloured beads with a micropipette and carefully transfer them to a small, clean Petri dish. 11 Add the 6.0 M guanidine HCl ph 1.0 solution to strip the protein off the bead. 12 Transfer the positive beads into a small clean Petri dish containing double distilled water. 13 Pipette each positive bead onto a glass filter and insert it into a protein sequencing cartridge for microsequencing. d a All the incubation and washing steps are performed in this column. b If there are too many positive beads, you should repeat the experiment by either decreasing the concentration of ligate alkaline phosphatase conjugate, or shortening the incubation time. c At this step, many colourless beads may be also transferred. d In some experiments one may want to decolorize the beads with DMF, and recycle the bead for another round of screening with an alternate assay method. With this approach, the probability for true-positive results will be significantly higher. 44

17 ONE-BEAD ONE-COMPOUND COMBINATORIAL LIBRARY METHOD Protocol 7 Unlabelled ligate detected with an enzyme-linked secondary antibody system Equipment and reagents Dissecting microscope Automatic protein sequencer (e.g. Applied Biosystems Model ABI 477A) Anti-ligate antibody/alkaline phosphatase conjugate Binding buffer: 16 mm Na 2 HPO 4, 3 mm KH 2 PO 4, 274 mm NaCl, 5.4 mm KCl ph 7.2, with 0.1% Tween (v/v) and 0.05% gelatin (w/v) Method 1 Pre-block 1 10 ml of the bead-supported library according to Protocol 6, step 1. 2 Wash the bead library thoroughly with 0.1% Tween/PBS (w/v). 3 Add a suitable concentration of the anti-ligate antibody/alkaline phosphatase conjugate to the bead-supported library and incubate for 3 h at room temperature. 4 Wash the bead-supported library thoroughly with PBS/Tween, and once with TBS. 5 Add the BCIP substrate buffer to the bead library and develop the colour as in Protocol 6, steps Remove all of the coloured beads with the aid of a micropipettte and a light box. 7 Transfer the colourless beads to the washing column. 8 Treat the colourless beads with 6.0 M guanidine HCl ph 1.0 for min, then wash with double distilled water (5 10 ml). 9 Mix the beads with DMF for 1 h for decolorization. 10 Wash serially with 10 ml each of 30% H 2 O/DMF, 60% H 2 O/DMF, and DMF. 11 Wash with double distilled water and 0.1% Tween/PBS (5 10 ml each). 12 Add the unlabelled ligate at a suitable concentration to the column containing recycled bead-supported library and incubate for 5 h at room temperature. 13 Wash the bead-supported library thoroughly with 0.1% Tween/PBS. 14 Add anti-ligate antibody/alkaline phosphatase conjugate (same concentration as step 3) to the bead-supported library for 2 h at room temperature. 15 Wash the bead-supported library thoroughly with 0.1% Tween/PBS and once with TBS. 16 Add BCIP substrate buffer to the bead library, transfer the bead-supported library to Petri dishes, and allow the colour reaction to develop as outlined in Protocol 6, steps 5 8. a 17 Stop the colour reaction by adding a few drops of 1.0 M HCl into the development buffer to adjust the ph to Isolate the positive beads for microsequencing as in Protocol 6, step 13. a The positive beads at this step are likely due to ligand ligate interaction rather than peptide/ anti-ligate antibody interaction since the latter interaction was in principle eliminated in the preceding steps. 45

18 GANG LIU AND KIT S. LAM Protocol 8 Cross-screening the library with enzyme-linked colorimetric and radiolabelled assays Equipment and reagents X-ray film (e.g. Kodak X-OMAT LS) Glogos II autoradiogram marker (Stratagene) Dissecting microscope Automatic protein sequencer (e.g. Applied Biosystems Model ABI 477A) Biotinylated ligate [ 125 I]ligate 1.0% low gelling temperature agarose (w/v) in H 2 O: melt the agarose in a microwave oven and keep at 37 C Method 1 Pre-mix the biotinylated ligate with streptavidin/alkaline phosphatase conjugate at a molar ratio of 4:1 a for at least 3 h at 4 C. 2 Treat the bead-supported library as in Protocol 6, steps 1 and 2. 3 Incubate the bead-supported library with the complex of biotinylated ligate and streptavidin/alkaline phosphatase conjugate in 0.05% gelatin/0.1% Tween /PBS at room temperature for 1 24 h. 4 Thoroughly wash the bead-supported library with 0.1% Tween/PBS (10 10 ml) and TBS (10 ml). Add BCIP as in Protocol 6, steps 5 8, then stop the colour development as in Protocol 6, step 9, and isolate the positive beads under the dissecting microscope. 5 Treat the positive beads with 6.0 M guanidine HCl ph 1.0, and then with DMF, as in Protocol 6, step Wash the decolorized beads with 0.1% Tween/PBS, followed by double distilled water. Block the decolorized beads with 0.1% gelatin again, as in Protocol 6, step 1. 7 Incubate the decolorized beads with streptavidin alkaline phosphatase alone at a concentration of 1:5000 for 2 h. 8 Wash the beads with 0.1% Tween/PBS (5 5 ml) and TBS (5 ml). Add BCIP for colour development as in Protocol 6, steps Remove and discard the coloured beads. 10 Transfer the remaining colourless beads into a 1 ml disposable polyethylene column, then wash the beads with double distilled H 2 O (5 5 ml), PBS/0.1% Tween (5 5 ml), and TBS (2 5 ml). 11 Incubate the beads with a suitable concentration of [ 125 I]ligate overnight at 4 C. 12 Wash the beads with 0.1% Tween in twice more concentrated PBS (10 5 ml). 13 Suspend the beads in 1.0% agarose. Carefully pour the bead suspension onto the clean glass plate and air dry it at room temperature. 14 Tape the Glogos II autoradiogram marker on the corner of the plate. 46

19 ONE-BEAD ONE-COMPOUND COMBINATORIAL LIBRARY METHOD Protocol 8 continued 15 Expose the immobilized beads to X-ray film overnight at room temperature. 16 Under the microscope, carefully pick up the positive beads that correspond to the dark spots of the film by excising the embedded bead into a polypropylene container with a small amount of water. 17 Heat it in the microwave to dissolve the agar. 18 Under the microscope, carefully transfer the beads into a Petri dish containing 6.0 M guanidine HCl ph 1.0 with the micropipette. 19 Transfer the positive beads to water and sequence the peptides as in Protocol 6, step 13. a Each streptavidin molecule has four binding sites. Protocol 9 Determination of peptide substrate motifs for protein kinases Equipment and reagents X-ray film (e.g. Kodak X-OMAT LS) Dissecting microscope Automatic protein sequencer (e.g. Applied Biosystem Model ABI 477A) MES buffer: 30 mm 2-(Nmorpholino)ethanesulfonic acid, 10 mm MgCl 2, 0.4 mg/ml bovine serum albumin (BSA) ph 6.8 [ - 32 P]ATP (25 Ci/mmol) Washing buffer: 0.68 M NaCl, 13 mm KCl, 40 mm Na 2 HPO 4, 7 mm KH 2 PO 4 ph 7.0, 0.1% Tween 20 (v/v) 0.1 M HCl 1.0% low gelling temperature agarose (w/v) in H 2 O a Method 1 Transfer 1 ml of bead-supported library into a disposable polypropylene column, wash the beads thoroughly with double distilled water, followed by MES buffer. 2 To the 1 ml settled bead volume add 1 ml of two times more concentrated MES buffer containing M [ - 32 P]ATP (adenosine triphosphate) and a specific protein kinase. Cap the reaction column tightly and put on a rocking platform for 1 5 h at room temperature with gentle rocking. 3 Wash the resin thoroughly with washing buffer, followed by double distilled water. 4 Transfer the 32 P-labelled bead-supported library to a glass container with 5 ml of 0.1 M HCl, and heat it to 100 C for 15 min to hydrolyse all the residual [ - 32 P]ATP. 5 Transfer the 32 P-labelled bead-supported library back into a clean disposable polypropylene column and wash the acid treated library with the washing buffer. 47

20 GANG LIU AND KIT S. LAM Protocol 9 continued 6 Suspend 0.5 ml of the 32 P-labelled bead-supported library resin in 30 ml of 1.0% agarose solution (70 C). Carefully pour the suspension onto a clean glass plate (16 18 cm), and air dry it overnight at room temperature. Tape the Glogos II autoradiogram markers on each corner of the glass plates. 7 Repeat step 6 for the remaining 0.5 ml bead library. 8 Expose the immobilized beads to X-ray film with an intensifying screen for h at room temperature and develop the film. 9 Align the autoradiograph with the Glogos II autoradiogram markers on the glass plates. Excise the area of the dried agar with the bead corresponding to the dark spots on the developed film. b 10 Collect all of the excised dried agar embedded beads in 30 ml of hot 1.0% agarose solution (70 C) for 15 min. 11 Replate the beads, expose, and develop the autoradiogram as described above. 12 Localize an individual bead that corresponds to the dark spot on the autoradiogram under the dissecting microscope, transfer the positive bead together with some attached agarose to a clean Petri dish containing a small amount of water. Heat it by microwave to dissolve the agar. 13 Under a dissecting microscope, transfer each positive bead onto a glass filter with a micropipette and insert into it the protein sequencer cartridge for microsequencing. a Melt agarose with microwave oven and keep at 70 C. b Generally, one dark spot corresponds to many beads. References 1. Lam, K. S., Salmon, S. E., Hersh, E. M., Hruby, V. J., Kazmierski, W. M., and Knapp, R. J. (1991). Nature, 354, Houghten, R. A., Pinilla, C., Blondelle, S. E., Appel, J. R., Dodey, C. T., and Cuervo, J. H. (1991). Nature, 354, Furka, A., Sebbstyen, F., Asgedom, M., and Dibo, G. (1991). Int. J. Pept. Protein Res., 37, Wu, J. Z., Ma, Q. N., and Lam, K. S. (1994). Biochemistry, 33, Salmon, S. E., Lam, K. S., Lebl, M., Kandola, A., Khattri, P., Wade, S., et al. (1993). Proc. Natl. Acad. Sci. USA, 90, Nikolaiev, V., Stierandova, A., Krchnak, V., Seligmann, B., and Lam, K. S. (1993). Peptide Res., 6, Ohlmeyer, M. H. J., Swanson, R. N., Dillard, L. W., Reader, J. C., Asouline, G., Kobayashi, R., et al. (1993). Proc. Natl. Acad. Sci. USA, 90, Brenner, S. and Lerner, R. A. (1992). Proc. Natl. Acad. Sci. USA, 89, Nicolaou, K. C., Xiao, X. Y., Parandoosh, Z., Senyei, A., and Nova, M. P. (1995). Angew. Chem. Int. Ed. Engl., 34, Geysen, H. M., Meloen, R. H., and Barteling, S. J. (1984). Proc. Natl. Acad. Sci. USA, 81,

21 ONE-BEAD ONE-COMPOUND COMBINATORIAL LIBRARY METHOD 11. Thompson, L. A. and Ellman, J. A. (1996). Chem. Rev., 96, Fruchtel, J. S. and Jung, G. (1996). Angew. Chem. Int. Ed. Engl., 35, Nefzi, A., Ostresh, J. M., and Houghten, R. A. (1997). Chem. Rev., 97, Lam, K. S., Lebl, M., and Krchnak, V. (1997). Chem. Rev., 97, Wahl, F. and Mutter, M. (1996). Tetrahedron Lett., 37(38), Spatola, A. F. and Romanovskis, P. (1996). Combinatorial peptide and nonpeptide libraries: a handbook, p 327. VCH Verlagsgesellschaft mbh, D Weinheim. 17. McMurray, J. S. (1994). Peptide Res., 7, Novabiochem Catalog and peptide synthesis handbook (1997/1998). p. S Lam, K. S. and Lebl, M. (1994). Methods: a companion to methods in enzymology, 6, Smith, M. H., Lam, K. S., Hersh, E. M., and Grimes, W. (1994). Mol. Immunol., 31, Pennington, M. E., Lam, K. S., and Cress, A. E. (1996). Mol. Div., 2, Lam, K. S., Wu, J. Z., and Lou, Q. (1995). Intl. J. Protein Peptide Res., 45, Lou, Q., Leftwich, M., and Lam, K. S. (1996). Bioorg. Med. Chem., 4, Meldal, M., Svendsen, I., Breddam, K., and Auzanneau, F. I. (1994). Proc. Natl. Acad. Sci. USA, 91, Meldal, M. and Svendsen, I. (1995). J. Chem. Soc. Perkin Trans., 1, Salmon, S. E., Liu-Stevens, R. H., Zhao, Y., Lebl, M., Krchnak, V., Wertman, K., et al. (1996). Mol. Div., 2, Lebl, M., Patek, M., Kocis, P., Krchnak, V., Hruby, V. J., Salmon, S. E., et al. (1993). Intl. J. Protein Peptide Res., 41,

22